NSF Grant for ML Modeling

We are thrilled to announce that FDT lab has received an NSF grant from the NSF Engineering Division to develop sea-state-dependent drag parameterizations through experiments and data-driven modeling. This collaborative project will integrate laboratory measurements of wind-wave interactions with a high-fidelity digital twin model of the laboratory system to develop a data-driven model for sea-surface drag. The specific objectives of the project are to understand skin friction modulations induced by surface waves, evaluate pressure drag through a high-fidelity digital twin model, and develop a sea-state-dependent total surface drag parameterization.

Laboratory measurements will accurately describe surface skin friction drag, but they fall short when it comes to pressure forces. The digital twin model will augment the experimental setup by providing pressure forces. This integrated approach will provide unique insight into wave-induced modulations of the total wind stress (sum of tangential and pressure stresses at the air-water interface) under a range of wind-wave conditions. A data-driven sea-state-dependent surface flux parameterization will be developed by examining these modulations, leveraging recent advancements in machine learning technology. The model will be tailored for large-eddy simulations of wind over ocean wavefields in strongly forced conditions. This approach is expected to significantly advance the fundamental understanding of air-sea fluxes and improve parameterizations of wind stress over the ocean.

Read more at National Science Foundation.